Gastric Cancer Single-Cell RNA-seq Analysis

This repository contains the code and data for the recreation and enrichment of the gastric (GC) cancer single-cell RNA-seq (scRNA-seq) data analysis pipeline described in the "Comprehensive analysis of metastatic gastric cancer tumour cells using single‑cell RNA‑seq" by Wang B. et. al, using the raw counts matrix that is available in GEO (GSE158631). Our robust and comprehensive pipeline surpasses previous approaches by incorporating multiple dimensionality reduction techniques, various clustering methods, marker gene identification and GO functional annotation.

This analysis was performed as part of the final project for the "Machine Learning in Computational Biology" graduate course of the MSc Data Science & Information Technologies Master's programme (Bioinformatics - Biomedical Data Science Specialization) of the Department of Informatics and Telecommunications department of the National and Kapodistrian University of Athens (NKUA), under the supervision of professor Elias Manolakos, in the academic year 2023-2024.

Contributors

• Konstantinos Giatras
• Olympia Tsiomou

Cloning the Repository

git clone https://github.com/GiatrasKon/Gastric-Cancer-scRNAseq-Analysis.git

Package Dependencies

Ensure you have the following packages installed:

• pandas
• IPython
• matplotlib
• seaborn
• numpy
• anndata
• scanpy
• scipy
• sklearn
• umap-learn
• gprofiler-official

Install dependencies using:

pip install pandas ipython matplotlib seaborn numpy anndata scanpy scipy scikit-learn umap-learn gprofiler-official

Class Description

The GCSingleCellAnalysis class (located in src/codebase.py) provides a comprehensive workflow for preprocessing, dimensionality reduction, clustering, and functional annotation of single-cell RNA-seq data.

Running the Analysis

1. Preprocessing and Normalization:

   from src.codebase import GCSingleCellAnalysis
   analysis = GCSingleCellAnalysis('data/GSE158631_count.csv')
   analysis.preprocess_adata()
   analysis.normalize_adata()
   analysis.filter_adata()
   analysis.prepare_adata()

2. Dimensionality Reduction:

   analysis.perform_pca()
   analysis.prepare_pca_reduced_adata()
   analysis.plot_pca()
   analysis.perform_tsne()
   analysis.plot_tsne()
   analysis.perform_umap()
   analysis.plot_umap()

3. Clustering:

   methods = ['gmm', 'average_link', 'ward', 'spectral', 'louvain', 'leiden']
   results_pca = analysis.cluster_and_evaluate(methods, embeddings=['X_pca'])
   results_tsne = analysis.cluster_and_evaluate(methods, embeddings=['X_tsne'])
   results_umap = analysis.cluster_and_evaluate(methods, embeddings=['X_umap'])
   combined_results = {'PCA': results_pca, 't-SNE': results_tsne, 'UMAP': results_umap}
   results_df = analysis.create_results_dataframe(combined_results)
   analysis.plot_clustering_evaluation(results_df)

4. Post-Clustering Analysis:

   analysis.analyze_and_plot_markers(group_key='average_link_X_umap', n_genes=10)
   go_annotations = analysis.fetch_go_annotations(group_key='average_link_X_umap', n_genes=10)
   analysis.print_go_annotations(go_annotations)

Notebook

For a detailed step-by-step analysis, refer to the Jupyter notebook:

• notebooks/GC_scRNAseq_data_analysis.ipynb

Images of the results of the analysis can be found in the images directory.

Documentation

Refer to the documents directory for the project proposal, presentation, report, and the original authors' paper.

---